The quantum efficiency of an image sensor is the ratio between the number of electronic charges collected and the number of photons incident on an active region of an image sensor, for example a photodiode. This quantity makes it possible to characterize the sensitivity of an image sensor to light.
The quantum efficiency for near-infrared optical signals, i.e. light rays whose wavelength lies between 700 nanometers and 1000 nanometers, is low, specifically because of the low absorption of these wavelengths by the silicon of the photosensitive regions of the sensors.
There are means which make it possible to improve the quantum efficiency of image sensors with backside illumination for near-infrared wavelengths.
One solution consists in increasing the optical path travelled by the light rays in the silicon by reflection phenomena, for example on metallization levels lying in the interconnection part (commonly referred to by the person skilled in the art by the acronym BEOL: “Back End Of Line”) behind the active zone, or on insulating trenches which delimit the active zone.
These methods, however, remain insufficient, particularly because of the small silicon thicknesses which are used in the most compact image sensors.
There is accordingly a need in the art to provide an image sensor having an improved quantum efficiency for infrared light rays.